The Node of Knowledge (NOK) conceptual framework describes a way of modeling natural language sentences into The Formalized Node of Knowledge (FNOK) record which in this form, can be used in a relational database. In addition to sentences, it also describes the method of modeling questions, i.e. interrogative sentences of natural language, into a formalized QFNOK record (Question Formalized Node of Knowledge). From the database defined and created in this way, it is possible to get answers to questions from natural language sentences. In previous research, a system was developed that enabled the recording of simple sentences and questions in natural language into the relational database of the Node of Knowledge (DBNOK), and it has been tested and applied to simple sentences in Croatian and English languages. The aim of this paper and research is to extend the existing system of writing simple natural language sentences to the relational database DBNOK with the possibility of writing complex English sentences. Complex sentences containing the conjunctions and/or and consisting of two actions or two process nodes, if viewed from the standpoint of the NOK conceptual framework are observed. The rules for transforming conjunctions into formalized FNOK and QFNOK records are defined in the paper with this aim. At the dictionary level, the system is extended by implementation of a solution to the problem of synonyms. These changes, in the formalized records of sentences and questions as well as the implementation of synonyms, also generated exstension in the system model, the relational database schema, algorithms for transforming formalized FNOK and QFNOK records into the relational database DBNOK and answer – seeking algorithms. The verification of the validity and functionality of all mentioned extension and their implementation was carried out by developing a system and verifying the validity of the answers that the system gives to the questions raised.